Power grid equipment supply chain risk identification method

Abstract

The invention discloses a power grid equipment supply chain risk identification method, and the method comprises the steps: determining the sample data related to risk identification through constructing a power grid equipment supply chain risk index system, building a power grid equipment supply chain risk identification model based on an incremental extreme learning machine of an attenuation regularization item, and completing the training of the model through employing the sample data. The power grid equipment supply chain risk identification model can accurately identify the current power grid equipment supply chain risk type of an enterprise, and the problems that in the prior art, many redundant hidden layer nodes reducing accuracy and learning efficiency exist, and hidden layer node parameters generated randomly affect the stability of an incremental extreme learning machine are solved. Requirements of power grid equipment supply chain risk identification can be met to a certain extent, and meanwhile, a new thought and a new approach are provided for more accurate power grid equipment supply chain risk identification.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a method for identifying risks in the supply chain of power grid equipment. Background Art

[0002] The power grid equipment supply chain is a core guarantee for the stable operation of power systems. However, its globalization and long-term nature make it susceptible to multiple risks, including supplier reliability, logistics efficiency, and market fluctuations. Existing research often uses methods such as the Analytic Hierarchy Process (AHP) and fuzzy comprehensive evaluation for risk assessment, but these methods have the following limitations: First, they are highly data-dependent, relying on sufficient historical data and struggling to address emerging risks or data-deficient scenarios; second, they lack dynamism, with existing static models unable to adapt to real-time changes in the supply chain; and third, they have weak anomaly detection capabilities and low sensitivity to sudden risks.

[0003] The existing incremental extreme learning machine (ELM) is a feedforward neural network learning algorithm. For single-hidden-layer feedforward neural networks, the connection weights between the input and hidden layers, as well as the thresholds of the hidden layer neurons, are randomly generated during training. Only the output weights from the hidden layer to the output layer need to be calculated. Consequently, existing prediction models based on traditional incremental ELMs contain many redundant nodes that reduce the accuracy and efficiency of risk identification in the power grid equipment supply chain. Randomly generated hidden layer node parameters can affect the stability of the ELM, leading to large network training errors. Therefore, designing efficient risk identification models has become a key issue in power grid equipment supply chain risk prediction research. Summary of the Invention

[0004] In view of this, the present invention provides a method for identifying risks in the power grid equipment supply chain. By introducing an attenuation regularization term in the training process of an incremental extreme learning machine model, an incremental extreme learning machine model based on the attenuation regularization term is constructed to achieve accurate identification of risks in the power grid equipment supply chain.

[0005] The present invention provides a method for identifying risks in the power grid equipment supply chain, which specifically includes the following steps:

[0006] Step 1: Collect risk indicator data related to the power grid equipment supply chain, construct an input matrix for the power grid equipment supply chain risk identification model, and form a training sample set; the risk indicator data includes quantitative indicator data on natural disaster risk, demand risk, product quality risk, inventory risk, financial risk, logistics risk, outsourcing risk, and cooperation risk;

[0007] Step 2: Construct a power grid equipment supply chain risk identification model based on an incremental extreme learning machine with a decay regularization term, and use the training sample set constructed in step 1 to complete the training of the power grid equipment supply chain risk identification model;

[0008] Step 3: During actual use, relevant quantitative indicator data of the power grid equipment supply chain risk indicator data within the enterprise to be evaluated is collected, and the collected quantitative indicator data is input into the trained power grid equipment supply chain risk identification model. The power grid equipment supply chain risk identification model outputs the risk type identification result of the power grid equipment supply chain currently faced by the enterprise to be evaluated.

[0009] Furthermore, the power grid equipment supply chain risk identification model is shown in the following formula:

[0010]

[0011] Among them, F 1×N ∈R 1×N is the output vector of the power grid equipment supply chain risk identification model, L is the total number of hidden layer nodes, is the output matrix of the i-th hidden layer node, W i 1×1 is the output weight of the i-th hidden layer node, α i is the attenuation regularization term corresponding to the i-th hidden layer node; i is a positive integer not greater than L, the initial value of i is 1 and satisfies 1≤i≤L; X 8×N ∈R 8×N is the input matrix of the power grid equipment supply chain risk identification model, is the input weight matrix between the nodes of the i-th hidden layer, is the bias matrix of the i-th hidden layer node; φ is an optional nonlinear activation function; and are all randomly generated and remain unchanged after generation; i represents the attenuation regularization coefficient corresponding to the i-th hidden layer node, e i Network residuals for supply chain risk identification models for power grid equipment; for The square of the 2-norm of ; for e i-1 The 2-norm of is raised to the power of .

[0012] Furthermore, a set of risk indicator systems for the power grid equipment supply chain is established, which can be expressed as:

[0013] X={x1,x2,x3,x4,x5,x6,x7,x8}

[0014] Among them, X is the supply chain risk indicator system; x1 is natural disaster risk, which mainly stems from force majeure factors in nature, such as earthquakes, floods, etc. that impact the supply chain; x2 is demand risk, which is mainly reflected in the uncertainty of electricity demand and the frequent changes in grid investment plans; x3 is product quality risk, which is related to the quality level of grid equipment and affects the stable operation of the grid; x4 is inventory risk, which involves the rationality and safety of inventory management; x5 is financial risk, covering aspects such as capital flow and cost control; x6 is logistics risk, including fluctuations in logistics costs and delivery punctuality; x7 is outsourcing risk, which mainly involves management and collaboration difficulties in the outsourcing business process; x8 is cooperation risk, which is related to the stability of cooperative relationships with suppliers and partners.

[0015] Furthermore, the input matrix is expressed as in, and are all matrices with 1 row and N columns, For natural disaster risk indicator data, Indicator data for demand risk, Indicator data for product quality risks, Indicator data for inventory risk, Indicator data for financial risk, Indicator data for logistics risks, Indicator data for outsourcing risks and is the indicator data of cooperation risk, and N is the total number of training samples of power grid equipment supply chain risk indicators.

[0016] Furthermore, the least squares method based on the attenuated regularization coefficient is used to obtain the output weights of the hidden layer nodes of the power grid equipment supply chain risk identification model.

[0017] Furthermore, the method for obtaining the output weights of the hidden layer nodes of the power grid equipment supply chain risk identification model is:

[0018] Step 1.1: Calculate the decay regularization coefficient corresponding to the i-th hidden layer node using the following formula:

[0019]

[0020] Among them, E 1×N is the error of the risk identification model for the power grid equipment supply chain, and its initial value is Y 1×N The ideal output vector for the risk identification model of power grid equipment supply chain;

[0021] Step 1.2: Calculate the output weight matrix of the i-th hidden layer node using the least squares method based on the attenuated regularization coefficient, as shown in the following formula:

[0022]

[0023] Where, is the error of the power grid equipment supply chain risk identification model when it contains i-1 hidden layer nodes; is the output matrix of the i-th hidden layer node; W i 1×1 is the output weight of the i-th hidden layer node; for and The inner product of

[0024] Step 1.3: Calculate the attenuation regularization term corresponding to the i-th hidden layer node using the following formula:

[0025]

[0026] Step 1.4: Based on the output weights and attenuation regularization terms calculated in steps 1.2 and 1.3, use the following formula to calculate the current error value of the power grid equipment supply chain risk identification model:

[0027]

[0028] Where, It represents the error of the power grid equipment supply chain risk identification model when it contains i hidden layer nodes;

[0029] Step 1.5: Let i add 1. When i≤L, execute step 1.1; otherwise, complete the training and output the output weight vector W of all hidden layer nodes. 1×L , ending this process.

[0030] Furthermore, after executing step 1.5, the following steps are further included:

[0031] Step 1.6: Construct the error between the output of the current power grid equipment supply chain risk identification model containing L hidden layer nodes and the ideal output vector of the power grid equipment supply chain risk identification model:

[0032]

[0033] in,

[0034] Step 1.7: Calculate the output weight deviation of the extreme learning machine based on the obtained error:

[0035]

[0036] in, is the output weight deviation of the extreme learning machine, is the generalized inverse matrix corresponding to the output matrix of the hidden layer nodes;

[0037] Step 1.8: Get the corrected solution of the extreme learning machine output weights:

[0038]

[0039] Furthermore, the natural disaster risk includes the frequency of historical disasters in the region, the proportion of affected facilities, the proportion of single disaster repair costs to the annual budget, and the number of days of supply chain disruption. Demand risk includes the monthly standard deviation of electricity demand, the percentage difference between actual demand and forecast values, the number of annual adjustments to the grid investment plan, and the lag time from policy adjustment to supply chain adjustment. Product quality risk includes the number of defects per thousand products, the pass rate of pre-shipment random inspections, the number of equipment failures in the first year of operation, and the time it takes to locate quality problems. Inventory risk includes the annual inventory turnover rate, the proportion of backlog materials to total inventory, the annual number of key material shortages, and the proportion of inventory management expenses to operating costs. Financial risk includes the monthly standard deviation of net cash flow, the average collection days, the annual price increase of raw materials, and the proportion of foreign currency settlement business. Logistics risk includes the proportion of deliveries on time, the quarterly fluctuation in logistics costs, the number of quarterly transportation delays, and carbon emissions per unit of cargo transportation. Outsourcing risk includes the proportion of outsourcers delivering contracts on time, the pass rate of outsourced products, the annual number of intellectual property disputes, and the proportion of business from a single outsourcer. Cooperation risk includes the number of years of cooperation with core suppliers, the proportion of real-time sharing of key data, the annual number of supplier contract breaches, and the response time for joint decision-making.

[0040] Furthermore, the training sample set is divided into two subsets for training and testing respectively.

[0041] Furthermore, the method of collecting risk indicator data related to the power grid equipment supply chain is to obtain it using big data analysis.

[0042] Beneficial effects:

[0043] The present invention determines sample data related to risk identification by constructing a power grid equipment supply chain risk indicator system, and then establishes a power grid equipment supply chain risk identification model based on an incremental extreme learning machine with a decaying regularization term. The sample data is used to complete the model training, so that the power grid equipment supply chain risk identification model can accurately identify the current power grid equipment supply chain risk type of the enterprise. It overcomes the problems in the existing technology of many redundant hidden layer nodes that reduce accuracy and learning efficiency, and the problem that randomly generated hidden layer node parameters affect the stability of the incremental extreme learning machine. It can meet the needs of power grid equipment supply chain risk identification to a certain extent, and at the same time provides new ideas and new approaches for more accurate power grid equipment supply chain risk identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of the processing flow of a power grid equipment supply chain risk identification method provided by the present invention.

[0045] Figure 2 A schematic diagram of a power grid equipment supply chain risk indicator system constructed by a power grid equipment supply chain risk identification method provided by the present invention. DETAILED DESCRIPTION

[0046] The present invention is described in detail below with reference to the accompanying drawings and with reference to the embodiments.

[0047] The present invention provides a method for identifying risks in the power grid equipment supply chain, the process of which is as follows: Figure 1 As shown, the specific steps include:

[0048] Step 1: Construct a set of risk indicator systems for the power grid equipment supply chain, expressed as:

[0049] X={x1,x2,x3,x4,x5,x6,x7,x8}

[0050] Among them, X is the supply chain risk indicator system, such as Figure 2 As shown in the figure; x1 is natural disaster risk, which mainly comes from force majeure factors in nature, such as earthquakes, floods and other impacts on the supply chain; x2 is demand risk, which is mainly reflected in the uncertainty of electricity demand and frequent changes in grid investment plans; x3 is product quality risk, which is related to the quality level of grid equipment and affects the stable operation of the grid; x4 is inventory risk, which involves the rationality and safety of inventory management; x5 is financial risk, covering aspects such as capital flow and cost control; x6 is logistics risk, including fluctuations in logistics costs and delivery punctuality; x7 is outsourcing risk, which mainly involves management and collaboration difficulties in the outsourcing business process; x8 is cooperation risk, which is related to the stability of the cooperative relationship with suppliers and partners.

[0051] The specific quantitative indicators of the power grid equipment supply chain risk indicator system are as follows:

[0052] Natural disaster risk (x1) includes: the frequency of historical disasters in the region (such as the average annual number of earthquakes and the ten-year recurrence period of floods), the proportion of affected facilities (such as the proportion of factories that have stopped production due to disasters), the proportion of single disaster repair costs to the annual budget (for example, the proportion of post-flood repair costs ≥15%), and the number of days of supply chain interruption (such as the average recovery period ≥30 days).

[0053] Demand risk (x2) includes: monthly standard deviation of electricity demand (such as ±20%), percentage difference between actual demand and forecast value (such as the proportion of months with deviation ≥10%), number of annual grid investment plan adjustments (such as ≥3 times / year), and lag time from policy adjustment to supply chain adjustment (such as an average of 60 days).

[0054] Product quality risks (x3) include: the number of defects per thousand products (such as ≤5 pieces / thousand pieces), the pass rate of random inspections before leaving the factory (such as ≥98%), the number of equipment failures in the first year of operation (such as ≤0.5 times / unit), and the time to locate quality problems (such as ≤24 hours).

[0055] Inventory risk (x4) includes: annual inventory turnover times (such as ≥6 times), the proportion of backlog materials in total inventory (such as ≤10%), the number of key material shortages / year (such as ≤2 times), and the proportion of inventory management expenses in operating costs (such as ≤8%).

[0056] Financial risk (x5): monthly net cash flow standard deviation (e.g. ±15%), average collection days (e.g. ≤60 days), annual raw material price increase (e.g. copper price increase ≤12%), proportion of foreign currency settlement business (if ≥20%, a warning is required).

[0057] Logistics risk (x6): Proportion of delivery on time (e.g. ≥95%), quarterly logistics cost fluctuation (e.g. ±10%), number of transportation delays / quarter (e.g. ≤3 times), carbon emissions per unit of cargo transportation (e.g. ≤50kg CO2 / ton·km).

[0058] Outsourcing risk (x7): proportion of contracts delivered on time by outsourcers (e.g. ≥90%), qualification rate of outsourced products (e.g. ≥95%), number of intellectual property disputes per year (e.g. ≤1), proportion of business with a single outsourcer (e.g. ≤30%).

[0059] Cooperation risk (x8): years of cooperation with core suppliers (e.g. ≥5 years), proportion of real-time sharing of key data (e.g. ≥80%), number of supplier contract breaches / year (e.g. ≤1 time), and joint decision-making response time (e.g. ≤48 hours).

[0060] Step 2: Use big data analysis to collect risk indicator data and build the input matrix of the power grid equipment supply chain risk identification model and are all matrices with 1 row and N columns, For natural disaster risk indicator data, Indicator data for demand risk, Indicator data for product quality risks, Indicator data for inventory risk, Indicator data for financial risk, Indicator data for logistics risks, Indicator data for outsourcing risks and is the indicator data of cooperation risk, and N is the total number of training samples of power grid equipment supply chain risk indicators, forming a training sample set.

[0061] The training sample set is divided into two parts, one for training and the other for testing. After the training of the power grid equipment supply chain risk identification model is completed, the test sample is used for testing. If the test result meets the preset requirements, the training is terminated, otherwise the training is executed again.

[0062] Step 3: Construct a power grid equipment supply chain risk identification model based on the incremental extreme learning machine with attenuation regularization term, as shown in formula (1):

[0063]

[0064] Among them, F 1×N ∈R 1×N is the output vector of the power grid equipment supply chain risk identification model, L is the total number of hidden layer nodes, is the output matrix of the i-th hidden layer node, W i 1×1 is the output weight of the i-th hidden layer node, α i is the attenuation regularization term corresponding to the i-th hidden layer node, i is a positive integer not greater than L (the initial value of i is i=1, and 1≤i≤L); X 8×N ∈R 8×N is the input matrix of the power grid equipment supply chain risk identification model, is the input weight matrix between the nodes of the i-th hidden layer, is the bias matrix of the i-th hidden layer node; φ is an optional nonlinear activation function; and are all randomly generated and remain unchanged after generation; i represents the attenuation regularization coefficient corresponding to the i-th hidden layer node, e i Network residuals for supply chain risk identification models for power grid equipment; for The square of the 2-norm of ; for e i-1 The 2-norm of is raised to the power of .

[0065] Step 4: Use the training sample set generated in step 2 to complete the training of the power grid equipment supply chain risk identification model established in step 3.

[0066] In order to further optimize the network structure, increase network stability, accelerate network convergence and improve computational efficiency, the present invention also proposes to use a least squares method based on a decaying regularization coefficient to obtain the output weights of the hidden layer nodes of the power grid equipment supply chain risk identification model, which specifically includes the following steps:

[0067] Definition, Y 1×N is the ideal output vector of the grid equipment supply chain risk identification model, E 1×N is the error of the risk identification model for the power grid equipment supply chain, and its initial value is

[0068] Step 1.1: Use formula (2) to calculate the attenuation regularization coefficient corresponding to the i-th hidden layer node:

[0069]

[0070] Step 1.2: Calculate the output weight matrix of the i-th hidden layer node using the least squares method based on the attenuation regularization coefficient, as shown in formula (3):

[0071]

[0072] Where, is the error of the power grid equipment supply chain risk identification model when it contains i-1 hidden layer nodes; is the output matrix of the i-th hidden layer node; W i 1×1 is the output weight of the i-th hidden layer node; for and The inner product of .

[0073] Step 1.3: Use formula (4) to calculate the attenuation regularization term corresponding to the i-th hidden layer node:

[0074]

[0075] Step 1.4: Based on the output weights and attenuation regularization terms calculated in steps 1.2 and 1.3, use formula (5) to calculate the current error value of the power grid equipment supply chain risk identification model:

[0076]

[0077] Where, It represents the error of the power grid equipment supply chain risk identification model when the power grid equipment supply chain risk identification model contains i hidden layer nodes.

[0078] Step 1.5: Let i add 1. When i≤L, execute step 1.1; otherwise, complete the training and output the output weight vector W of all hidden layer nodes. 1×L , ending this process.

[0079] In order to further improve the accuracy of the training results of the power grid equipment supply chain risk identification model, the present invention adds the following steps after completing steps 1.1 to 1.5:

[0080] Step 1.6: Construct the error between the output of the current power grid equipment supply chain risk identification model containing L hidden layer nodes and the ideal output vector of the power grid equipment supply chain risk identification model:

[0081]

[0082] in,

[0083] Step 1.7: Calculate the output weight deviation of the extreme learning machine based on the error obtained from formula (6):

[0084]

[0085] in, is the output weight deviation of the extreme learning machine, is the generalized inverse matrix corresponding to the output matrix of the hidden layer nodes.

[0086] Step 1.8: Obtain the corrected solution of the extreme learning machine output weights according to formulas (6) and (7):

[0087]

[0088] Step 5. During actual use, the enterprise's internal power grid equipment supply chain data is collected, including a certain number of supply chain risk quantitative indicator data, including natural disaster risk, demand risk, product quality risk, inventory risk, financial risk, outsourcing risk, cooperation risk and logistics risk; the collected quantitative indicator data is input into the trained power grid equipment supply chain risk identification model, and the power grid equipment supply chain risk identification model outputs the risk type identification result of the power grid equipment supply chain currently faced by the enterprise.

[0089] Implementation Cases:

[0090] This embodiment adopts a power grid equipment supply chain risk identification method provided by the present invention, and uses an incremental extreme learning machine model based on a decaying regularization term to identify power grid equipment supply chain risks. The method specifically includes the following steps:

[0091] Create a training sample set The data comes from power grid equipment supply chain data collected internally by enterprises, including 90 supply chain risk information samples, covering disaster risk, demand risk, quality risk, inventory risk, financial risk, outsourcing risk, cooperation risk, and logistics risk. i represents the i-th supply chain risk information sample, y i The value is 0 or 1, y i The position k with a value of 1 indicates that the information sample belongs to the kth category, and the value of 0 indicates that it does not belong to the corresponding category.

[0092] The false recognition rate and recognition rate are used as model evaluation indicators. The false recognition rate and recognition rate are important evaluation indicators for measuring the accuracy of the recognition method. The false recognition rate is the ratio of the number of incorrectly recognized samples to the total number of samples. The smaller the value, the less the recognition method misclassifies the samples, and the higher the recognition accuracy. The recognition rate is the ratio of the number of correctly recognized samples to the total number of recognized samples. The larger the value, the stronger the ability of the recognition method to accurately identify samples and the higher the recognition accuracy. These two indicators reflect the performance of the recognition method from different angles and play a key role in evaluating the accuracy of the power grid equipment supply chain risk identification model. The specific calculation formula is as follows:

[0093]

[0094] In the formula, FAR stands for the false positive rate (FAR) of risk in the power material supply chain, reflecting the proportion of incorrectly identified samples throughout the risk identification process. TF represents the number of normal samples incorrectly identified as risky, TN represents the number of risky samples incorrectly identified as normal, and NF represents the number of correctly identified samples. FFR stands for the risk identification rate in the power material supply chain, reflecting the proportion of correctly identified samples in the total number of identified samples. RF represents the total number of identified samples.

[0095] The recognition method based on BP wavelet neural network proposed in reference [1], the recognition method based on blockchain technology proposed in reference [2], and the recognition method based on generative adversarial network proposed in reference [3] were selected for comparison with the method in this paper. The experimental results are shown in Tables 1 and 2.

[0096] Table 1. Risk misidentification rate of power grid equipment supply chain

[0097]

[0098] Table 2 Risk identification rate of power grid equipment supply chain

[0099]

[0100] Analyzing the data in Tables 1 and 2, we conclude that among the three methods, the blockchain-based identification method has the highest false positive rate, followed by the BP neural network-based method. The proposed method has the lowest false positive rate, at only 0.14%. In terms of recognition rate, our proposed method achieves a recognition rate exceeding 95%. The blockchain-based method, on the other hand, achieves a maximum recognition rate of only 81.36%, and the BP neural network-based method achieves a maximum recognition rate of only 77.52%, both significantly lower than our proposed method. This comparison and analysis demonstrates that our proposed method offers significant advantages in accurately identifying risks in the power grid equipment supply chain.

[0101] Finally, it should be noted that it is apparent that those skilled in the art may make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, to the extent such modifications and variations fall within the scope of the claims and their equivalents, the present invention is intended to encompass such modifications and variations.

[0102] The above is only an embodiment of the present invention, but it is not intended to limit the scope of the present invention. Any structural changes made according to the present invention, as long as they do not lose the essence of the present invention, should be considered to fall within the scope of protection of the present invention and be subject to restrictions. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working process and related instructions of the method described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0103] The term "comprise," "comprising," or any other similar term is intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus / method that comprises a list of elements includes not only those elements but also other elements not expressly listed or inherent to such process, method, article, or apparatus / method.

[0104] Thus far, further embodiments have been listed to describe the technical solutions of the present invention. However, it is readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

[0105] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for identifying risks in the power grid equipment supply chain, characterized in that: The specific steps include: Step 1: Collect risk indicator data related to the power grid equipment supply chain, construct an input matrix for the power grid equipment supply chain risk identification model, and form a training sample set; the risk indicator data includes quantitative indicator data on natural disaster risk, demand risk, product quality risk, inventory risk, financial risk, logistics risk, outsourcing risk, and cooperation risk; Step 2: Construct a power grid equipment supply chain risk identification model based on an incremental extreme learning machine with a decay regularization term, and use the training sample set constructed in step 1 to complete the training of the power grid equipment supply chain risk identification model; Step 3: During actual use, relevant quantitative indicator data of the power grid equipment supply chain risk indicator data within the enterprise to be evaluated is collected, and the collected quantitative indicator data is input into the trained power grid equipment supply chain risk identification model. The power grid equipment supply chain risk identification model outputs the risk type identification result of the power grid equipment supply chain currently faced by the enterprise to be evaluated.

2. A method for identifying risks in the power grid equipment supply chain according to claim 1, characterized in that: The power grid equipment supply chain risk identification model is shown in the following formula: Among them, F 1×N ∈R 1×N is the output vector of the power grid equipment supply chain risk identification model, L is the total number of hidden layer nodes, is the output matrix of the i-th hidden layer node, is the output weight of the i-th hidden layer node, α i is the attenuation regularization term corresponding to the i-th hidden layer node; i is a positive integer not greater than L, the initial value of i is 1 and satisfies 1≤i≤L; X 8×N ∈R 8×N is the input matrix of the power grid equipment supply chain risk identification model, is the input weight matrix between the nodes of the i-th hidden layer, is the bias matrix of the i-th hidden layer node; φ is an optional nonlinear activation function; and are all randomly generated and remain unchanged after generation; i represents the attenuation regularization coefficient corresponding to the i-th hidden layer node, e i Network residuals for supply chain risk identification models for power grid equipment; for The square of the 2-norm of ; for e i-1 The 2-norm of is raised to the power of .

3. A method for identifying risks in the power grid equipment supply chain according to claim 1, characterized in that: A set of risk indicator systems for the power grid equipment supply chain is established, which can be expressed as: X={x1,x2,x3,x4,x5,x6,x7,x8} Among them, X is the supply chain risk indicator system; x1 is natural disaster risk, which mainly stems from force majeure factors in nature, such as earthquakes, floods, etc. that impact the supply chain; x2 is demand risk, which is mainly reflected in the uncertainty of electricity demand and the frequent changes in grid investment plans; x3 is product quality risk, which is related to the quality level of grid equipment and affects the stable operation of the grid; x4 is inventory risk, which involves the rationality and safety of inventory management; x5 is financial risk, covering aspects such as capital flow and cost control; x6 is logistics risk, including fluctuations in logistics costs and delivery punctuality; x7 is outsourcing risk, which mainly involves management and collaboration difficulties in the outsourcing business process; x8 is cooperation risk, which is related to the stability of cooperative relationships with suppliers and partners.

4. A method for identifying risks in a power grid equipment supply chain according to claim 1, characterized in that: The input matrix is represented as in, and are all matrices with 1 row and N columns, For natural disaster risk indicator data, Indicator data for demand risk, Indicator data for product quality risks, Indicator data for inventory risk, Indicator data for financial risk, Indicator data for logistics risks, Indicator data for outsourcing risks and is the indicator data of cooperation risk, and N is the total number of training samples of power grid equipment supply chain risk indicators.

5. A method for identifying risks in a power grid equipment supply chain according to claim 1, characterized in that: The least squares method based on the attenuated regularization coefficient is used to obtain the output weights of the hidden layer nodes of the power grid equipment supply chain risk identification model.

6. A method for identifying risks in the power grid equipment supply chain according to claim 5, characterized in that: The method for obtaining the output weights of the hidden layer nodes of the power grid equipment supply chain risk identification model is: Step 1.1: Calculate the decay regularization coefficient corresponding to the i-th hidden layer node using the following formula: Among them, E 1×N is the error of the risk identification model for the power grid equipment supply chain, and its initial value is Y 1×N The ideal output vector for the risk identification model of power grid equipment supply chain; Step 1.2: Calculate the output weight matrix of the i-th hidden layer node using the least squares method based on the attenuated regularization coefficient, as shown in the following formula: Where, is the error of the power grid equipment supply chain risk identification model when it contains i-1 hidden layer nodes; is the output matrix of the i-th hidden layer node; is the output weight of the i-th hidden layer node; for and The inner product of Step 1.3: Calculate the attenuation regularization term corresponding to the i-th hidden layer node using the following formula: Step 1.4: Based on the output weights and attenuation regularization terms calculated in steps 1.2 and 1.3, use the following formula to calculate the current error value of the power grid equipment supply chain risk identification model: Where, It represents the error of the power grid equipment supply chain risk identification model when it contains i hidden layer nodes; Step 1.5: Let i add 1. When i≤L, execute step 1.1; otherwise, complete the training and output the output weight vector W of all hidden layer nodes. 1×L , ending this process.

7. A method for identifying risks in the power grid equipment supply chain according to claim 6, characterized in that: After executing step 1.5, also include: Step 1.6: Construct the error between the output of the current power grid equipment supply chain risk identification model containing L hidden layer nodes and the ideal output vector of the power grid equipment supply chain risk identification model: in, Step 1.7: Calculate the output weight deviation of the extreme learning machine based on the obtained error: in, is the output weight deviation of the extreme learning machine, is the generalized inverse matrix corresponding to the output matrix of the hidden layer nodes; Step 1.8: Get the corrected solution of the extreme learning machine output weights:

8. A method for identifying risks in a power grid equipment supply chain according to claim 1, characterized in that: Natural disaster risks include the frequency of historical disasters in the region, the proportion of affected facilities, the proportion of single-disaster repair costs to the annual budget, and the number of days of supply chain disruption. Demand risk includes the monthly standard deviation of electricity demand, the percentage difference between actual demand and forecast values, the number of annual adjustments to the grid investment plan, and the lag time from policy adjustments to supply chain adjustments. Product quality risk includes the number of defects per thousand products, the pass rate of pre-shipment random inspections, the number of equipment failures in the first year of operation, and the time it takes to locate quality issues. Inventory risk includes the annual inventory turnover rate, the proportion of overstocked materials to total inventory, the annual number of key material shortages, and the proportion of inventory management expenses to operating costs. Financial risk includes the monthly standard deviation of net cash flow, the average collection days, the annual price increase of raw materials, and the proportion of foreign currency settlement business. Logistics risk includes the proportion of deliveries on time, the quarterly fluctuation in logistics costs, the number of quarterly transportation delays, and carbon emissions per unit of cargo transportation. Outsourcing risk includes the proportion of outsourcers delivering contracts on time, the pass rate of outsourced products, the annual number of intellectual property disputes, and the proportion of business with a single outsourcer. Cooperation risk includes the number of years of cooperation with core suppliers, the proportion of real-time sharing of key data, the annual number of supplier contract breaches, and the response time for joint decision-making.

9. A method for identifying risks in a power grid equipment supply chain according to claim 1, characterized in that: The training sample set is divided into two subsets for training and testing respectively.

10. A method for identifying risks in a power grid equipment supply chain according to claim 1, characterized in that: The method of collecting risk indicator data related to the power grid equipment supply chain is to obtain it using big data analysis.

Images